An individual node may initially operate a database independently of other nodes in a local only mode of operation. At a later point in time, the individual node may join a group of nodes and begin operating in a distributed mode. In distributed mode, data of the individual node's database may be merged with data of databases of other nodes. However, the transition from local only mode to distributed mode may lead to inefficiencies. For example, the transition may include copying data, which may be error prone and may waste resources, such as memory and processing resources. Furthermore, operations started in the individual node during the local only mode may be negatively affected by the transition. For example, the operations may try to use invalid data, may be stopped, or may have to be restarted.